Exome Sequencing of Sporadic GH-Secreting Pituitary Adenomas

Growth hormone pituitary adenomas (GHomas) are the second most common of pituitary tumors that hypersecrete prolactin. In clinically, the treatment and diagnosis of GHomas are limited, and the earlier and more frequent recurrences were demonstrated in invasive pituitary adenomasa. The identification of the association of genetic mutation and GHomas progression can provide significant interest to patients with GHomas. In recent years, studies of GHomas focused on the identification of genetic mutation in sporadic pituitary adenomas. Previously, a study has shown DPCR1, EGFL7, the PRDM family and LRRC5 in pituitary adenomas are probably functional in modifiers of tumorigenesis, development of oncocytic change and invasive tumor phenotype. The detection of genetic mutations still needs further study. Whole-exome sequencing, for spectrum of genetic alterations at present study, was performed on 7 sporadic GHomas DNA and corresponding blood samples. In total, 293 variants were predicted to be functionally damaged in 7 tumor samples (median 41 mutations/sample; range: 40-73).


Introduction
The incidence of pituitary adenomas in the population is approximately 1: 1000 [1], which is 15% of all intracranial neoplasms [2]. Growth hormone adenoma is an extremely common functional pituitary adenoma, second only to pituitary tumors that hypersecrete prolactin. Only 5% of pituitary adenomas occur in patients with a familial history [1] and most are sporadic Growth Hormone (GH)-secreting pituitary adenomas [3], which cause acromegaly in adults, and excessive GH secretion during childhood/ adolescence can lead to gigantism [4]. Pituitary adenomas are thought to be monoclonal in origin [5], but the genesis of pituitary tumors has remained controversial. These tumors can cause serious complications, including ophthalmological, neurological and endocrinological abnormalities. Currently, factors that increase tumor progression remain uncharacterized, although many of the genetic changes associated with familial pituitary adenomas, including Multiple Endocrine Neoplasia type 1 (MEN1), Carney's complex and MEN4 [6][7][8], as well as germline mutations in the aryl hydrocarbon receptor interacting protein (AIP) gene and X-chromosomal microduplication, can predispose individuals to pituitary adenomas [6,9].
Somatic mutations in GNAS have been identified in 30-40% of GH-producing pituitary adenomas [7]. Interestingly, 53% of Japanese patients with GH-secreting pituitary adenomas have been reported to exhibit somatic GNAS mutations [8]. Mutations in the gene that encodes the α subunit of stimulatory G-protein, Gs, are the only mutational changes unequivocally associated with GH-secreting adenomas [10]. Pathological analyses have identified proliferation markers that were unaltered in mutated GNAS pituitary tumors and non-mutated tumors, suggesting that the GNAS1 mutation affected secretion rather than proliferation [11]. Only 5% of pituitary adenoma patients have a familial history and despite multiple extensive studies, no oncogene or tumor suppressor genes have been discovered to directly affect the pathogenesis of sporadic GH-secreting pituitary adenomas. In order to discover genes involved in these adenomas, whole-exosome sequencing can be utilized to search for new somatic mutations in sporadic GH-secreting pituitary adenomas. Recently, Ronchi et al. examined adenoma tissue genomes by targeted sequencing (n = 31) and next-generation exome-sequencing (n = 36). No recurrent somatic mutations were observed, except for known alterations in the GNAS gene, which was like results collected by the wholegenome sequencing (n = 12) performed by Valimaki et al. [13,14].
Together, this indicates that pathogenesis in sporadic GH-secreting pituitary adenomas remains to be clarified. In order to further explore the somatic landscape of sporadic GH-secreting pituitary adenomas, we performed whole-exome sequencing coupled to rigorous analytical methods, which further confirmed novel recurrent genetic alterations and important mutated genes.  Note: Abbreviation: GH, Growth Hormone. Estimated tumor volume using the Di Chiro-Nelson method of V=1/2(h*w*l) as documented on preoperative magnetic resonance imaging.

Whole-Exome Sequencing and Bioinformatics Analysis
Exome enrichment was performed using the Agilent Sure Select Human All Exon V5+ UTR, while sequencing was performed with the Illumina Hiseq 4000 platform outputting 150-bp reads. Sequencing data were aligned to hg19/GRCh37 using the burrows-wheel aligner (bwa v0.7.12). Single nucleotide variants and deletions were identified using mutect-1.1.7 and strelka-1.0.11. Candidate driver mutations were identified using MutSigCV-1.4.

Gene Ontology and Pathway Enrichment Analysis for Mutated Genes
Gene Ontology analysis (GO) is a common method for annotating genes and gene products, while also for identifying characteristic

Integration of Protein-Protein Interaction (PPI) Network and Module Analysis
The    Note: The number of Somatic DNA mutations detected in 7 growth hormone adenoma sample. SNVs was detected by mutect and indel was detected strelka. Total mutations mean all variation found in whole exome sequencing region. Novel mutation mean this mutation is not in dbSNP. Missense, nonsense, splice-site, UTR, synonymous, non-coding mean which function variation has.

Whole-Exome Sequencing
Heterozygous and homozygous mean whether this variation is diploid.

Identification of Candidate Driver Mutations
Consistent with the absence of a family history for pituitary adenomas, the tumor samples lacked germline mutations in genes encoding MEN1, CDKN1B, AIP, and PRKAR1A, which are familial genes in pituitary adenomas. In addition, no mutations were found in known oncogenes, tumor suppressor genes, or genes previously implicated in other pituitary tumors. In order to identify candidate driver mutations, MutSigCV-1.4 was used to correct for variations by using patient-specific mutation frequency and spectrum, as well as gene-specific background mutation rates that were incorporated into expression level and replication time [15]. Using these criteria, 4 candidate driver mutations genes were identified: RBM43, KRTAP4-9, GNGT2 and CENPW. Previously, only GNGT2 had been implicated in tumorigenesis and epigenetic changes.

Gene Ontology and Pathway Analyses
Using GO analyses, genes were significantly enriched in metabolic processes and biological regulation (Biological Processes), the gene products were concentrated in the membrane and nucleus (Cellular Component), and the functions related to protein and ion binding (Molecular Function) (Figure 2). Using the KEGG database pathway analysis tool, there were no significant overlaps, however the discovered genes were associated with metabolic, calcium signaling, and adipocytokine signaling pathways Figures S1-S4), Yet there was not significant enrichment of the pathways and, integrating the data with the PPI network and module analysis, none of the genes were recognized as hub-genes. genes [6,13,14,17,18]. 4 candidate driver mutations were observed in this study: RBM43, KRTAP4-9, GNGT2, and CENPW. Only GNGT2 variants have been associated with any diseases. GNGT2, localized to chromosome 17q21, encodes the guanine nucleotide binding protein, Gamma Transducing Activity Polypeptide 2 (GNGT2) [19], which is thought to play a crucial role in cone phototransduction.
However, GNGT2 is important in obesity, interacting with the chemokine signaling pathway to affect the risk of pancreatic cancer [20] and methylation of GNGT2 is linked to the tobacco smoking to risk of coronary artery disease [21].  [12]. Further studies are needed to examine the functional relevance of this pathway.
Intriguingly, The GO term analysis showed that mutations genes were mainly involved developmental biological processes and ion transport in cancer cells differs substantially from normal cells [29]. Furthermore, the enriched KEGG pathways of mutations genes included calcium signaling, metabolic and adipocytokine signaling pathways, which indicates that these pathways are related, although there was no significant enrichment of these pathways.
Activated Ca 2+ signaling leads to an increase in cytosolic free calcium, which then further triggers GH secretion [30,31]. These findings agree with a recent study that described whole-genome alterations in 12 GH-secreting tissues using next-generation exomesequencing (n = 36) [13,14]. Other pathways that participate in pituitary adenomas have not been reported, although some heterogeneous tumors have been observed to share a few common mutations [14]. Together, these data indicate that different genes act through the same molecular pathways and may contribute to tumor formation in sporadic GH-secreting pituitary adenomas.

Conclusion
We did not identify any novel or recurrent mutated genes from GH-secreting pituitary adenomas. However, these genetic findings, coupled with previous studies, suggests that different genes act through similar pathways or epigenetic changes, which may contribute to tumor occurrence and development.

Funding Source
This project was supported by the national natural science foundation of China (81772685), the international cooperation research projects of Shenzhen Science and Technology Program (GJHZ20160301163419476 and GJHZ20160301163900284).